Azure Advenced Analytics Platform for Machine Learning
Transcript of Azure Advenced Analytics Platform for Machine Learning
Azure Advenced Analytics Platform for Machine LearningElitmind for Business
www.elitmind.com
EN
Solution description
Based on many years of experience, we've builta ready-made solution based on the scaledAzure cloud platform for advanced analytics. Itenables our client to exit from the standardmany separate solutions into one, used forbuilding and managing Machine Learning and AImodels.
READY-TO-GO PRODUCT FOR MACHINE LEARNING & AI
Solution products
• Development of a scalable environment architecture foradvanced analytics (ML / AI).
• Implementing the solution by making a Pilot of the selectedMachine Learning model
• As part of the implementation, we offer:• data potential analysis• data preparation and purification• selection, development and testing of the model
implementation and alignment (or support at the sametime
• For an additional fee, we can automate the entireenvironment in terms of integration with domain systems.
READY-TO-GO PRODUCT FOR MACHINE LEARNING & AI
• Customer segmentation – what are the types of customers• Anti-churn – which customers are likely to leave and how
to prevent it• Recommenders – what should be recommended to the
customers based on their purchasing history• Sales forecasting – how much will be sold• Price forecasting – how much will energy, derivatives or
products cost• Reviews analysis – what is being said about products
The most popular ML modelsREADY-TO-GO PRODUCT FOR MACHINE LEARNING & AI
Architecture diagramREADY-TO-GO PRODUCT FOR MACHINE LEARNING & AI
Data loading and orchestration
Azure Data Factory + Azure Automation
Blob Storage / Data Lake Storage
IP
Power BI
Database systems
E-commerce
Cloud solutions
Corporate data
Machine Learning &TransformationData integration
layer
Azure Databricks
IP
IP
Code management and versioning. Production
sharing.
Azure DevOps / ML OPS
Azure EventHub & Stream Analytics
(optional)
Azure MLStudio
IPEnd business user
Real –time streamdata analysis services.
AzureKubernetes
Services
Azure SQL Database
Deploy andoperationalize a ML
model as API
Results and daat forreporting
Reporting & analytics
IP
Azure price simulationREADY-TO-GO PRODUCT FOR MACHINE LEARNING & AI
Product components Details Monthly price $ for up to 10 users
O365 Power BI Pro 9,99 $ per User 99,9 $
Azure services (middle size) ADFv2, ADLv2, ADB/AMLS, ASQLDB ~3 000 $
Total ~3 100 $
Product components Details Monthly price $ for up to 10 users
O365 Power BI Pro 9,99 $ per User 99,9 $
Azure services (large size) ADFv2, ADLv2, ADB, AMLS, ASQLDB, AKS ~5 000 $
Total ~5 100 $
Azure componentsREADY-TO-GO PRODUCT FOR MACHINE LEARNING & AI
Cloud services interacting with
IP
Cloud services that are hosting
the IP
Data conn., data layers etc.
Security control etc.
User interfaces On-premises conn. /
integration
• Event Hub• Stream
Analytics• AKS• Power BI• SQLDatabase
• Blob storage / Data Lake Storage
• Databricks• Machine
Learning Service
• ML Ops• Data Factory
• Data Factory• Azure
Automation• VPN Gateway• Integration
Runtime
• Data Factory• VPN Gateway• Integration
Runtime
• Power BI• AKS
• VPN Gateway• Integration
Runtime
Fast & Agile
▪ Scrum methodology with Azure DevOps
▪ Faster and frequent delivery of products to customer by integrative delivery
▪ Less formalism by using tools and developed practices
Reliable Team of Experts
▪ We deliver what we promised
▪ Our team is individually selected to meet the needs of the project
▪ 95% of consultants Microsoft certified
▪ 9 years of experience on average
Quality and Partnership
▪ We always think strategically and build long term partnerships
▪ We use our frameworks to make projectspredictable and measurable
▪ We use experience to provide the highest quality standards
How We Do
Price & implementation approachREADY-TO-GO PRODUCT FOR MACHINE LEARNING & AI
• The implementation can be on a dedicated customer Azure subscription or an existing Elitmind Azure subscription (CSP). If it is not possible to separate a dedicated subscription, it is possible to use a dedicated Resource Group.
• The implementation price includes:• Fixed fee for Azure services (Azure fees statement)• Preliminary analysis & architecture described in the document with
recommendations: 3 to 6 days (3 000,0 $)• A fee determined after the initial analysis & architecture for adapting
the finished elements to the client's requirements (2 to 6 weeksimplementation).
• Additional fees for T&E.
Data science capabilities
http://elitmind.com/services/consulting/#Advanced-analytics
• Data Mining: Big Data exploration including classification, prediction and clustering
• Text Mining: text patter recognition, social media and email analysis.
• Forecasting: statistical modelling and econometric techniques for forecasting purpose.
• Optimization: state-of-the art economic methods, advance statistical analysis and mathematical optimization.
• Data science in Organization: processes, competences, data potential, priorities & ROI, awareness and inspiration.
DATA GATHERING & STORAGE
Marketing & Customer AnalyticsCASE STUDY: Analytics Platform for Marketing
Retail & E-commerce
Problem: manual and error-free process ofhandling and preparing marketing campaigns.
Solution: automatic process of data integrationfrom business systems and reporting campaigneffectiveness in near real time, thematic datawarehouse, predictive models for customersegmentation and churn analysis, integrationwith marketing automation.
Business Value: time saving, elimination oferrors in reports, possibility of immediatereaction to problems in campaigns.
IP-COSELL IMPLEMENTATION
Logistic & Customer AnalyticsCASE STUDY: Anti-Churn Model
Problem: uncontrolled process of outflow ofbusiness customers.
Solution: Azure Machine Learning hybrid model,based on boosted methods, black-box model -indicating risk-generating factors. 3 months project.
Business value: Increasing revenue by anticipatingthe possibility of quitting the services of a businessclient.
An international logistics company
IP-COSELL IMPLEMENTATION
Energy companyDerivativesCASE STUDY: ENERGY RESOURCES SPOT PRICES FORECASTING
Problem: no easy accesible data for makingdecisions on energy resources derivativestransactions
Solution: analytical environment and predictionmodel for spot prices, hybrid model combiningshort-term approach with focus on accuracy andlong-term with focus on interpretenabilty
Business value: better investment decisions basedon reports of spot prices predictions, less than 5 %error on average in short term
IP-COSELL IMPLEMENTATION
About Elitmind
We BUILD the world of data for our Partners
• We save time and money by automating data processing.
• We increase company efficiency through tailor-made analyzes and reports and by adopting self-service analytics.
• We build competitive advantage by predicting the future based on the past.
Selected Customers
Phone: +48 501 626 668
Phone: +48 605 269 855Phone: +48 505 191 532
Digital Advisors & Founders
Chief Technology Officer
https://www.elitmind.com
https://www.linkedin.com/company/elitmind-sp--z-o-o-/
https://twitter.com/elitmind
https://www.facebook.com/elitmind/
https://www.elitmindacademy.com